İstatistik Bölümü Tez Koleksiyonuhttp://hdl.handle.net/11655/3002021-03-09T02:23:22Z2021-03-09T02:23:22ZTIP BİLİŞİMİNDE VERİ MADENCİLİĞİ YÖNTEMLERİ KULLANILARAK HASTALIKLARIN TAHMİN EDİLMESİSıtkı, Yasemin Handehttp://hdl.handle.net/11655/232922021-01-22T09:14:02Z2020-01-01T00:00:00ZTIP BİLİŞİMİNDE VERİ MADENCİLİĞİ YÖNTEMLERİ KULLANILARAK HASTALIKLARIN TAHMİN EDİLMESİ
Sıtkı, Yasemin Hande
Nowadays, Data Mining is an increasingly important tool, especially in the administration of healthcare enterprises and in determining health-related policies, as it provides support for the decision-making processes of businesses by revealing information hidden from large dimensional data. Moreover, scientific publications have been made in the field of health in recent years to diagnose diseases using data mining algorithms.
In this thesis, most frequently used data mining classification methods were examined, and a study was conducted to diagnose 18 different diseaes in the urology branch using the data collected from the patients who applied to the urology branch of four different public hospitals. For this purpose, classification algorithms Random Forest, Random Tree, Multilayer Perception, IBk, Kstar one of the sample based algorithms, Simple Logistic and Naive Bayes from statistical algorithms and ZeroR from rule learning algorithms were used, and the correct classification rates of the created models, namely how correctly they diagnosed the diseases, were examined.
Among these algorithms Random Forest, Simple Logistic and Multilayer Perception algorithms have been found to be more successful in diagnosis than others. In future studies, an application can be developed for the diagnosis of diseases related to urology branch or diseases seen other branches by using the algorithms mentioned here. Thus, it may be possible to give healthcare professionals an idea in the diagnosis and to reduce their workload, to find the diseases in advance with early diagnosis and to shorten the treatment period.
2020-01-01T00:00:00ZEwma Control Charts For Skewed DıstrıbutıonsMoustapha, Amınou Tukurhttp://hdl.handle.net/11655/232172021-01-07T12:38:54Z2020-09-30T00:00:00ZEwma Control Charts For Skewed Dıstrıbutıons
Moustapha, Amınou Tukur
The classic Shewhart control charts are generally used for monitoring the process mean and variability in the characteristics of a random quality variable of interest and are based on the normality assumptions. For skewed distributions, in order to demonstrate the changes in the population, non-symmetric control limits need to be used. Methods such as the Weighted Variance (WV) Weighted Standard Deviation (WSD) and Skewness Correction (SC) are used with skewed distributions.
The classic 𝑋� ̅ and R control charts and all their derivatives are generally used to detect large shifts in the process mean hence making them not too reliable in situations where in small shifts are of interest. To solve such problems, the Exponentially Weighted Moving Average (EWMA) control charts is used in this work.
The main aim of this thesis is to apply the Skewness Correction method to the EWMA chart and propose a control limit called Skewness Correction EWMA (SC-EWMA) for skewed distributions. The performances of the newly proposed method are compared and contrasted with those of the Weighted Variance EWMA (WV-EWMA) which was developed by Khoo and Atta (2008), Weighted Standard Deviation EWMA (WSD-EWMA) which was developed by Atta and Ramli (2011) and the classic EWMA control limits based on the degree of skewness and varying smoothing parameters. The comparison is made with respect to their type-Ι errors by using the Monte Carlo simulation technique with data generated from the lognormal, Gamma and Weibull distributions.
2020-09-30T00:00:00ZBayesci grafik modelleriTuğrul, Özgürhttp://hdl.handle.net/11655/228092020-11-05T08:14:34Z2001-01-01T00:00:00ZBayesci grafik modelleri
Tuğrul, Özgür
Posterior distributions obtained by Bayesian approach usually have high dimensions when the models are rather complicated. Therefore, to reach the marjinal distributions from the models is analitically intractable. The aim of the study is to introduce the newest Bayesian practical techniques for the complicated models and to investigate Bayesian graphical model and Gibbs Sampling which is a Markov Chain Monte Carlo method together. BUGS package is used to show how the complicated models are solved iteratively. To understand the conditional structure of the models is very crucial for analysing the complicated models. Graphical models can be used to have a visual information of the structure of these models. Conditional independence allows us to factorize the joint distributions. Samples can be drawn iteratively from the marjinal distributions of the model parameters by Markov Chain Monte Carlo techniques. Then statistical inference can be make easier for the marjinal distiributions. Three different models are investigated in the study. BUGS package is used to have the visual representations of the three models. Gibbs samplings is applied for these models to obtain the marjinal distributions of the model parameters. The result obtained from Gibbs samplings are compared with the classical results in the multiple regression model. The changes in the number of iterations, initial values and the prior distributions of the models parameters are investigated. When the number of the iteration increases, the results are very close to true values. It is also seen in the study that the initial values are not so important if the number of iterations is high.
2001-01-01T00:00:00ZYinelemeli Sinir Ağları ile Finansal Veri TahminiKeçeci, Ekinhttp://hdl.handle.net/11655/227892020-12-10T09:00:16Z2020-08-01T00:00:00ZYinelemeli Sinir Ağları ile Finansal Veri Tahmini
Keçeci, Ekin
Recurrent Neural Network is an artificial neural network model which the outputs are re-included to network input in every iteration. The biggest advantage of recurrent neural networks is that they consider the variation of each sample in the sequential data depending on the previous examples. As reccurent neural network models developed, some theoretical obstacles emerged and different models were developed as solutions to these obstacles. It can be said that Long term short term memory (LSTM) networks are one of the most popular and best designed reccurent neural network models among these models. In this study, the success of LSTM model is compared to other neural network models in evaluating financial asset prices. The LSTM model gave better classification accuracy than other compared models.
2020-08-01T00:00:00Z